Phase I/II Trial of Single Session Stereotactic Body Radiotherapy for Previously Un-irradiated Spinal Metastases

Abstract

Background

This Phase I/II study tests the hypothesis that single-fraction SBRT for previously un-irradiated spinal metastases is a safe, feasible, and efficacious treatment approach.

Methods

All patients were evaluated by a multidisciplinary team. Spinal MRI was performed before treatment and at regular intervals to both define target volume and response to treatment. SBRT was delivered to a peripheral dose of 16–24 Gy in 1 fraction while limiting dose to the spinal cord. Higher doses were used for renal cell histology. The NCI Common Toxicity Criteria 2.0 and McCormick neurological function score were used as toxicity assessment tools.

Results

A total of 61 patients harboring 63 tumors of the non-cervical spine were enrolled and treated with SBRT between 2005 and 2010 on a prospective Phase I/II trial at the University of Texas M. D. Anderson Cancer Center. Mean follow-up was 20 months. Actuarial 18-month imaging local control for all patients was 88%. Actuarial 18-month overall survival for all patients was 64%. Median survival for all patients was 30 months. No significant differences in outcomes were noted with respect to tumor histology and SBRT dose. Two patients experienced radiation adverse events (Grade 3 or higher). Actuarial 18-month freedom from neurologic deterioration from any cause as was 82%.

Conclusions

This Phase I/II data support an expanded indication for SBRT as first-line treatment of selected spinal metastases patients. Additional studies that can prospectively identify predictive factors for spinal cord toxicity after SBRT are warranted to minimize the incidence of this serious yet rare complication.

Introduction

Stereotactic body radiotherapy (SBRT) allows the precise delivery of tumor ablative radiation doses to tumors and high-risk post-operative regions while maximizing avoidance of nearby critical structures. Our group and others have recently shown that this technique is both safe and effective in achieving durable local tumor control as well as maintaining neurologic function in spinal metastases patients where spinal cord toxicity is of utmost concern.^1–4

Conventional fractionated radiotherapy for the treatment of spinal metastases is well-established, however is limited in its ability to deliver tumor ablative doses to tumors that are near the spinal canal. While many patients may achieve a favorable palliative response to conventional therapy, progression of disease is becoming a growing concern with improvements in patient survival. We have recently shown that patients with progressive spinal disease after conventional treatment may benefit from reirradiation with SBRT, however progressive and recurrent spinal disease is often associated with pain and worsened performance status and quality of life.^{2, 5} Spinal tumors close to the cord, regardless of previous therapy, can progress to cause spinal cord compression, a complication frequently associated with rapid clinical deterioration and poor prognosis if not able to be surgically decompressed.^{6, 7} Therefore, strategies to enhance local tumor control in the upfront treatment of spinal tumors may ultimately improve patient quality of life, reduce the need for salvage therapies, and ultimately improve outcomes in select patients with limited comorbidities.

The optimal dose and fractionation pattern of SBRT for spinal lesions remains unknown, however recent reports have demonstrated that a variety of schedules may be effective.^{4, 5, 8} While fractionated SBRT remains an attractive treatment option for patients with spinal disease, the ability to deliver SBRT in a single dose is more appealing if considered safe and effective. There may also be radiobiologic advantages. Herein we report, to our knowledge, the largest prospective phase I/II trial to date evaluating single-fraction SBRT in previously un-irradiated non-cervical spinal lesions.

Methods

Patients with a history of prior radiotherapy to the spine at the current level of interest, or chemotherapy within 30 days were excluded. Patients with acute spinal cord compression or spinal instability were not eligible. All patients had a Karnofsky performance status of at least 40. All cases were evaluated and discussed in a multidisciplinary spinal tumor board including members from neurosurgery, radiation oncology (including research nurses), and radiation physics.

At simulation, patients were immobilized in an Elekta BodyFix stereotactic body frame system and aligned using a stereotactic localizer and target positioning frame (Integra-Radionics, Burlington, Mass). Intrathecal injection of contrast (Omnipaque, Amersham Health, Bothell, Wash) was used in cases where imaging alone was not sufficient to clearly delineate the spinal cord (as in the case of spinal instrumentation or proximity of tumor to spinal cord).

Gross tumor volume (GTV) was defined as gross tumor on MRI and CTV was defined as GTV plus contiguous marrow space or at-risk postoperative region. Planning treatment volume was defined as the clinical treatment volume with no additional expansion for PTV secondary to precision of stereotactic setup technique.

Treatment planning was performed using IMRT inverse-treatment planning software (Pinnacle version 6.2, Phillips Medical Systems, Andover, Mass). Patients were stratified according to histology (renal versus non-renal) as well as location (spinal versus paraspinal) for purposes of dose prescription. For non-renal spinal metastases (thoracic, lumbar, or sacral spine), the mean gross tumor volume (GTV) was generally prescribed to receive 18 Gy and the mean clinical target volume (CTV) received 16 Gy. Spinal tumors with renal-cell histology received 24 Gy to the GTV and 16 Gy to the CTV. For paraspinal tumors, non-renal cell histologies generally received 18 Gy to the GTV and renal-cell histology received 24 Gy to the GTV. In general, 80–90% of the target volume received the prescription dose.

In general, dose to the spinal cord was limited to no greater than 0.01 cc of the spinal cord itself receiving greater than 10 Gy based on tabular dose-volume histogram (DVH) analysis and the spinal cord + 2mm limited to 12 Gy or less, with no greater than 0.01cc of the spinal cord itself receiving greater than 10 Gy based on tabular dose-volume histogram (DVH) analysis. However, final plan approval and spinal cord dose was at the discretion of the treating physician and final prescription dose to the target volume was reduced to meet spinal cord constraints if deemed necessary by the treating physician, such as in cases where there is epidural disease or disease close to the spinal cord (≤ 1mm).

A detailed dosimetric analysis was performed for all patients and the following variables were recorded: distance of tumor to the spinal cord (< 5mm), maximal point dose to the cord (D_max), maximal dose to 0.01 cc (D_0.01cc), 0.1 cc (D_0.1cc), and 1 cc (D_1cc) of the spinal cord, and maximal dose to 10% of the spinal cord contour as defined by 6 mm above and below the level of the tumor (D_10%). Additionally these volumetric variables were re-calculated for the thecal sac, as defined by the spinal cord + 1.5 mm expansion according to Sahgal et al.⁹

Spinal MRI was performed as a baseline before treatment initiation and at pre-defined follow-up intervals (3-month intervals for the first year and then every 6 months). Local failure was defined by MRI-documented progression of the treated spinal or paraspinal tumor as determined by the expert opinion of a dedicated radiologist. Patients were assessed for toxicity using the National Cancer Institute Common Toxicity Criteria, version 2.0 and neurologic function was assessed using the McCormick scale.¹⁰ Pain levels specific to the level of spine undergoing stereotactic body radiotherapy treatment were evaluated using the Brief Pain Inventory at baseline before stereotactic body radiotherapy, and at defined follow-up intervals.

All patients underwent intensity-modulated, near-simultaneous, CT-guided stereotactic body radiotherapy as previously described. ^{1, 2, 11} Near-simultaneous pretreatment CT scan was performed using the Trilogy cone-beam CT system (Varian Medical Systems, Palo Alto, Calif) and 3D-3D matching was performed. Verification of target positioning and quality assurance procedures for each case was performed by the radiation oncologist and dedicated radiation physicists, respectively. All patients received SBRT in 1 fraction.

Statistical analysis, including actuarial analysis of tumor progression and survival, and a t test between means for pain levels were generated using Stata 10 statistical software (StataCorp, College Station, Tex). Time to event curves were calculated from the date of registration on the study protocol. Clopper-Pearson analysis was used to generate early stopping criteria for the trial which included the evidence of paralysis (Grade 4 motor neuropathy) caused by radiation myelitis and not by tumor progression or spinal cord compression at a rate greater than 0.01 (1%).¹²

Results

Between 2005 and 2010, 61 patients with 63 individual spinal metastases of the thoracic, lumbar or sacral spine were enrolled on a prospective Phase I/II trial at the University of Texas M. D. Anderson Cancer Center designed to evaluate the safety and efficacy of using SBRT to treat non-cervical spine and paraspinal tumors in a single session. The patient and tumor characteristics for all patients are presented in Table I. The mean follow-up time for all patients was 19.7 months (range, 1.2–52.1 months; median, 17.8 months). Of the study cohort, 55% (34 of 62) patients were alive at the time of analysis. There were 8 local spinal failures in 63 total tumors treated with single fraction stereotactic body radiotherapy. Tumor histology and pertinent dosimetric characteristics for these 8 patients are shown in Table II. Of all 63 tumors, dose reduction was deemed necessary in 10 patients to improve spinal cord dose; however no patient received less than 16 Gy in a single fraction. Of the 8 patients with progressive disease, 2 patients required a moderate dose reduction to achieve dose constraints imposed upon the cord (1 patient with breast cancer received 16 Gy instead of 18 Gy and 1 patient with renal cell cancer received 20 Gy instead of 24 Gy). Total dose to the tumor (≥ 20 Gy vs ≤ 18 Gy) did not correlate with any clinical endpoint assessed.

Table I.

Patient and Tumor Characteristics for all patients

Median Age (range)	60 (34–78)
Gender
Male	34
Female	27
Number of Treated Targets Per Patient
1	59
2	2
Spinal Region Treated
Thoracic	39
Lumbar	23
Sacral	1
Karnofsky Performance Scale
100	6
90	19
80	23
70	10
60	3
Prior Surgery Type
Corpectomy	13
Other	3
None	47
Prior Radiation to Treated Spinal Level
Yes	0
No	63
Tumor Histology
Renal	33
Thyroid	10
Sarcoma	6
Breast	5
Lung	3
Others	6
Median Volumes (range, cm3)
Gross Tumor Volume (GTV)	25.64 (0.21–99.23)
Clinical Target Volume (CTV)	51.67 (3.89–210.5)
Planning Target Volume (PTV)	58.19 (3.89–227.63)
Neurologic Function at Baseline (McCormick)
1	54
2	2
3	5

Table II.

Tumor and Dosimetric Characteristics of the 8 patients With Local Progression

Patient	Primary Histology	Cord Distance1	Dose/Fractions2	Cord Dmax3	Tumor V804
1	Breast	> 5mm	16 Gy/1	888 cGy	>95%
2	Thyroid	< 5mm	18 Gy/1	1180 cGy	>99%
3	Renal Cell	> 5mm	24 Gy/1	1236 cGy	100%
4	Thyroid	> 5mm	18 Gy/1	1220 cGy	100%
5	Renal Cell	< 5mm	24 Gy/1	1388 cGy*cauda	100%
6	Renal Cell	post-op	20 Gy/1	1487 cGy*cauda	100%
7	Sarcoma	< 5mm	18 Gy/1	1258 cGy	98%
8	Endometrial	post-op	18 Gy/1	1209 cGy	100%

¹Distance of tumor to cord on initial MRI

²Total dose/# fractions

³Maximum dose to the spinal cord

⁴Percentage of gross tumor volume (GTV) receiving 80% of the prescription dose

^*Maximum dose to cauda as tumor located below spinal cord at L3

Spinal Tumor Control

The actuarial 18-month local tumor control for all patients was 88% (Fig. 1). Actuarial 18-month rates of local tumor control in patients treated postoperatively (n = 16) was 100%, with only 1 postoperative patient (6%) with imaging evidence of local progression 32 months after treatment. Patients who achieved durable pain control (< 4 out of 10) at 3 months and 6 months from the date of their treatment had better actuarial local tumor control when compared to patients with worse pain control (P ≤ 0.01, Fig. 2a/b: 1-year local control rate of 100% vs 84% for pain control at 3 months and 1-year local control rate of 97% vs 87% for pain control at 6 months; 2-year local control rate of 100% vs 77% for pain control at 3 months and 2-year local control rate of 97% vs 77% for pain control at 6 months). The Kaplan-Meier estimates of tumor progression did not correlate with other treatment factors including postoperative status (P = 0.32), dose (≥ 20 Gy, P = 0.99), primary tumor histology (renal vs non-renal, P = 0.67), spinal cord distance from tumor (≥ 5 mm vs < 5 mm, P = 0.58), and tumor volume (≥ 50 cm³ vs < 50 cm³, P = 0.85).

Figure 1.

Local tumor control for all patients

Figure 2.

Local tumor control according to pain level

a. At 3 months from treatment

b. At 6 months from treatment

A univariate Cox regression analysis of patient characteristics with respect to tumor control was performed using age cutoff (age < 60 years, P = 0.10), Karnofsky performance score (≤ 80, P = 0.73), stereotactic body radiotherapy dose (≤ 18 Gy, P = 0.99), primary tumor histology (renal vs non-renal, P = 0.45; renal vs thyroid, P = 0.31; renal vs sarcoma, P = 0.62), postoperative status (P = 0.42), tumor volume (≥ 30 cm³, P = 0.57; ≥ 50 cm³, P = 0.85), neurologic function preservation (yes vs no, P = 0.80), spinal level (thoracic vs lumbar/sacral, P = 0.18), pain free status at baseline (yes vs no, P = 0.33), and durable pain control (<4 out of 10 at 6 months, P = 0.06). On multivariate analysis, failure to achieve durable pain control <4 out of 10 at 6 months was the only factor independently associated with worse local tumor control (P = 0.04, hazard ratio 9.4).

Overall Survival

The median survival time for all study patients was 30.4 months. The actuarial 18-month survival was 64% (Fig. 3). For the 16 patients treated postoperatively, overall survival was not significantly different than survival of patients who did not undergo surgery prior to SBRT (P = 0.31). Patients who achieved durable pain control (< 4 out of 10) at 3 months from the date of their treatment had better overall survival when compared to patients with worse pain control (100% vs 77% at 1 year and 77% vs 52% at 2 years, P = 0.004, Fig. 4). The Kaplan-Meier estimates of overall survival did not correlate with other treatment factors including postoperative status (P = 0.31), primary tumor histology (renal vs non-renal, P = 0.22), dose (≥ 20 Gy, P = 0.23), spinal cord distance from tumor (≥ 5 mm vs < 5 mm, P = 0.87), and tumor volume (≥ 50 cm³ vs < 50 cm³, P = 0.66).

Figure 3.

Overall survival for all patients

Figure 4.

Overall survival according to pain level at 3 months from treatment

A univariate Cox regression analysis of patient characteristics with respect to overall survival was performed using age cutoff (age < 60 years, P = 0.26), Karnofsky performance score (≤ 80, P = 0.06), stereotactic body radiotherapy dose (≤ 18 Gy, P = 0.24), primary tumor histology (renal vs non-renal, P = 0.22; renal vs thyroid, P = 0.52; renal vs sarcoma, P = 0.90), postoperative status (P = 0.48), tumor volume (≥ 30 cm³, P = 0.53; ≥ 50 cm³, P = 0.66), neurologic function preservation (yes vs no, P < 0.01), spinal level (thoracic vs lumbar/sacral, P = 0.91), pain free status at baseline (yes vs no, P = 0.77), and durable pain control (<4 out of 10 at 3 months, P = 0.01 and at 6 months P = 0.23). On multivariate analysis, failure to achieve durable pain control <4 out of 10 at 3 months (P = 0.04, hazard ratio 3.28) and 6 months (P = 0.03, hazard ratio 5.28) from the time of treatment independently predicted for worse overall survival.

Preservation of Neurologic Function

The actuarial freedom from neurologic deterioration according to grade of McCormick neurologic function score was 82% at 18 months (Fig. 5). There were a total of 8 of 61 (13%) patients who experienced deterioration in neurologic function. No patients had neurologic function deterioration as a result of progressive local disease at the treated site. Three of 61 (4.9%) patients had neurologic deterioration due to progressive disease outside of the treated region at 2–2.5 years after SBRT to the index lesion. Two of 61 (3%) patients had neurologic deterioration attributed to radiation toxicity from stereotactic radiotherapy to the index lesion, the first patient developed hemicord syndrome 11 months after treatment for metastatic breast cancer at T1 manifesting as myelopathy with bilateral lower extremity weakness and left foot drop, and the second patient with an L5 radiculopathy arising 9 months from treatment for a renal cell metastasis manifesting as right foot drop with associated numbness and pain. The patient with an L5 radiculopathy remained ambulatory with a cane and required pain medication and the patient with myelopathy was unable to ambulate and required a wheelchair. Finally, 3 of 61 (4.9%) patients had neurologic function deterioration due to other causes, including severe peripheral neuropathy attributed to systemic chemotherapy in 1 patient, bilateral progressive rising paraplegia and eventual paralysis attributed to Guillain-Barre syndrome in 1 patient, and progressive right lower extremity weakness and pain attributed to a crush injury in 1 patient. There was 1 patient who developed malignant epidural spinal cord compression due to progressive disease approximately 6 months after SBRT at T5 treated successfully with salvage surgery with no loss of neurologic function.

Figure 5.

Neurologic progression-free survival

Radiation Toxicity

The radiation treatment-related toxicity is summarized in Table III. Two patients experienced radiation adverse events (Grade 3 and Grade 4 neurologic toxicity as described above). Cases of mild toxicity (Grade 1 or 2) included neurologic toxicity attributable to transient numbness and tingling (10 patients), gastrointestinal toxicity manifesting as transient nausea and vomiting (6 patients) or esophagitis/reflux (5 patients), and other toxicity most commonly fatigue (9 patients). There were no cases of Grade 4 toxicity. Thirteen patients had evidence of vertebral fracture (Grade 1 or 2 musculoskeletal toxicity) on follow-up imaging and were either treated with kyphoplasty, vertebroplasty, or no further intervention if thought to be stable.

Table III.

Treatment Toxicity

	Neurotoxicity	Hematologic	Gastrointestinal	Musculoskeletal	Other toxicity (worst grade)
None	50	61	49	48	47
Grade 1	6	0	9	7	9
Grade 2	3	0	3	6	5
Grade 3	1	0	0	0	0
Grade 4	1	0	0	0	0

Pain Relief

A greater number of patients were pain free at 3 months and 6 months (n = 18) when compared to those who were pain free at baseline (n =13), and as a group experienced reduced pain levels (≤ 3 vs ≥ 4) at 3 months and 6 months when compared with pain levels recorded at baseline; however, these results were not statistically significant (P ≥ 0.3).

Dosimetric Analysis

A comparison of mean dose values and standard deviations for all dosimetric variables for the spinal cord and thecal sac is shown in Figure 6. Mean D_max for spinal cord and thecal sac were 1200 cGy (range 776 – 1821 cGy) and 1449 cGy (range 904 – 2100 cGy), respectively. Mean tumor V80 was 99.5% (range 94–100%). None of the variables assessed for the spinal cord or thecal sac (D_max, D_0.01cc, D_0.1cc, D_1cc, D_10%) correlated with local tumor progression or neurologic function drop (P > 0.07).

Figure 6.

Dosimetric variables for spinal cord and thecal sac

a. Spinal Cord^{1, 2}

b. Thecal Sac^{1, 2,3}

¹Whisker box-plots representing median values (vertical bar) within box encompassing 25^th and 75^th percentile values and whiskers extending to 5^th and 95^th percentile values.

²D_max = maximal point dose, D_0.01cc = maximal dose to 0.01 cc, D_0.1cc = maximal dose to 0.1 cc, D_1cc = maximal dose to 1 cc, D_10% = maximal dose to 10% of contoured volume as defined by 6 mm above and below level of tumor.

³Thecal sac is defined by spinal cord volume + 1.5 mm

Discussion

To our knowledge, this is the largest prospective series of previously un-irradiated spinal disease treated with single-fraction stereotactic body radiotherapy. With a high rate of local tumor control (88% at 18 months) and limited toxicity (3% with Grade 3 or higher neurotoxicity). our results support an expanded indication for the first-line use of single-fraction SBRT to prescribed doses as high as 24 Gy. A steep dose gradient that is achievable with advanced treatment planning systems and adequate immobilization allows tumor ablative doses to be delivered to gross disease and microscopic disease in high-risk postoperative regions while respecting dose limits imposed on nearby spinal cord and other normal tissues that would often receive a greater proportion of the prescribed dose if conventional fields were to be used.^{13, 14} A favorable median survival time of 30 months in this cohort also suggests that properly selected patients may enjoy enhanced quality of life with durable control of spinal disease that may be more likely to progress after lower dose conventional radiotherapy. These include patients with “radioresistant” histologies like renal cell carcinoma and patients with limited systemic disease who may be expected to live longer and benefit from continually improving systemic and salvage therapies.^{3, 8} Tumor ablative doses using stereotactic body radiotherapy may also effectively prevent or reduce the incidence of spinal cord compression, as demonstrated in our study by the limited incidence of spinal cord compression for tumors within 5 mm of the spinal cord (1 of 17, 6%). Finally, single fraction SBRT significantly reduces total treatment time (1 fraction versus the standard multi-week palliative regimens) allowing more timely initiation of additional therapies if needed.

The results of this prospective study are consistent with the favorable results of recently reported retrospective data. The Memorial Sloan-Kettering Cancer Center (MSKCC) reported an actuarial local control rate of 90% and limited toxicity (no Grade 3) in 103 spinal tumors treated to 18–24 Gy with single-fraction SBRT.¹⁵ Additionally, in a separate report of 126 non-spinal extracranial tumors treated with single-fraction SBRT, the MSKCC group found renal cell histology to be profoundly dose-responsive, with 81% local control for renal cell tumors receiving 23–24 Gy versus 30% local control for renal cell tumors receiving ≤ 22 Gy (P = 0.03).¹⁶ Since our study found no such difference in local control rates among different histologies (P = 0.88), it may be suggested that our initial stratification of renal cell tumors to receive higher doses (24 Gy) adequately compensated for its more radioresistant behavior. This is also supported by recent data in the postoperative setting, where patients with highly radioresistant histologies achieved high rates of local tumor control (>90% at 1 year, n = 21) after decompressive surgery and adjuvant high-dose single-fraction SBRT. ¹⁷ When taken together with the results of our current study (2-year postoperative local control of 100%, n = 16), high-risk postoperative patients may also expect favorable outcomes with single-dose SBRT.

Conventional fractionated radiotherapy for metastatic spinal tumors has long been a standard of care and is characterized by its ease of setup and treatment planning; however the current study provides strong evidence that upfront single-fraction SBRT may be a more favorable approach for the initial management of spinal tumors in select patients who may benefit from more durable local tumor control.^{5, 18, 19} The total dose delivered to the spine using conventional radiotherapy is limited by the perceived tolerance of the spinal cord to fractionated radiotherapy (i.e. an upper limit of 45 _Gy2/2), thus preventing the delivery of more definitive doses.^{2, 5} Recent reviews of radiotherapy for spinal disease confirm that response to conventional radiotherapy using a wide variety of dose and fractionation schedules is often limited to under a year.^{5, 19, 20} Moreover, in the setting of spinal cord compression, a randomized trial demonstrated that upfront conventional radiotherapy (30 Gy in 10 fractions) resulted in worse clinical outcome as measured by ambulatory status over time when compared to patients treated with initial surgery followed by radiotherapy, suggesting that conventional radiotherapy dose was insufficient in controlling the disease.⁷

While patients with clinical characteristics of spinal cord compression were excluded from the current study, a significant number of patients included in our trial had tumors that were within 5 mm of the spinal cord (n = 17), with only 1 patient within that subset (6%) developing spinal cord compression due to progressive disease approximately 6 months after receiving 18 Gy in a single fraction for metastatic sarcoma to the T5 spinal level. The patient subsequently underwent salvage surgery without loss of neurologic function and lived pain-free before dying of systemic disease 6 months later. These results are to be contrasted with the results from a recently reported prospective study by our group that used a similar setup and technique to reirradiate spinal tumors and high-risk postoperative regions with hypofractioned SBRT (27 Gy in 3 fractions or 30 Gy in 5 fractions) after receiving prior radiotherapy.² In that study, the majority of spinal tumors that progressed after reirradiation with hypofractionated SBRT (13 of 16, 81%) were within 5 mm of the spinal cord with 6 of those 13 (46%) patients eventually developing malignant spinal cord compression. More stringent spinal cord dose constraints used to permit reirradiation were likely to be responsible, at least in part, for this pattern of failure.

In aggregate these findings indicate a significant dose-response effect of gross disease near the cord, whereby tumors treated in the salvage setting received far less radiobiologic dose in order to respect the nearby spinal cord restriction of 9–10 Gy in 3 to 5 fractions versus a more lenient spinal dose restriction of 10–12 Gy in 1 fraction with up-front single fraction SBRT in the current study. The results may also suggest that lower dose conventional radiotherapy may select for more aggressive and radioresistant disease in patients who live long enough to experience a recurrence, a hypothesis described previously in other tumor sites.²¹ What remains to be seen is whether high-dose single-fraction SBRT in the setting of previously irradiated spinal tumors is feasible, however concerns for spinal cord toxicity would make further dose-escalation a daunting proposal, especially with already high local control rates (>75%) and limited toxicity (<4% Grade 3 or 4 neurotoxicity) following salvage reirradiation with fractionated SBRT as described by our group and others.^{2, 22} Of note, guidelines for spinal cord tolerance in the re-irradiation setting have recently been published.²³

An incidence of Grade 3 or 4 neurotoxicity less than 4% (2 patients) in our study would suggest that current planning systems and careful delivery are adequate in generating dose gradients steep enough to achieve acceptable doses to the spinal cord without compromising local tumor control. In a detailed dosimetric analysis of each patient in our study, we found that point doses to the spinal cord often reached 12 Gy with limited toxicity, a result consistent with other experiences.^{9, 23, 24} Of the 2 patients who developed ≥ Grade 3 neurotoxicity, the first developed an L5 radiculopathy after receiving 24 Gy to his metastatic renal cell carcinoma with a maximal nerve root point dose of 1467 cGy and the second patient developed a radiation-induced myelopathy (Figure 7) after receiving 18 Gy to her metastatic breast cancer at T1 with a maximal spinal cord point dose of 1267 cGy. Both patients remained without local progression of their treated spinal disease until the first patient died of progressive systemic disease 17 months after SBRT. The second patient remains alive 24 months after SBRT (at last follow-up) and requires a wheelchair.

Figure 7.

Patient with radiation-induced myelopathy

a. CT-based stereotactic treatment plan demonstrating isodose lines and contoured structures including gross tumor (red), clinical target volume (blue), and spinal cord (purple).

b. T2-weighted MRI demonstrating abnormal left cord hyperintensity (yellow arrow) consistent with radiation-induced myelopathy.

When interpreting spinal cord tolerance data, it should be noted that differences in treatment planning systems, dose grid sizes, spinal cord delineation methods (i.e. use of myelogram), and contour definitions and uncertainties (spinal cord vs thecal sac, spatial inaccuracies with fused MRI images) are among many important variables that should be considered when comparing different studies. Nonetheless, our results indicate a favorable therapeutic ratio for the use of SBRT in spinal tumors and support a growing consensus that traditional models of radiation-induced cell killing are inadequate in accounting for the clinical response seen with hypofractionated radiotherapy.^25–28 An important caveat is that SBRT is a high-risk, high-reward procedure that should be performed at experienced centers with adequate resources, and not widely adopted in the community without adequate training and resources.

In conclusion, the current prospective study using a homogenous technique of SBRT delivery supports an expanded indication for the use of single-fraction SBRT in the treatment of previously unirradiated spinal tumors and in high-risk postoperative patients. Patients can be expected to achieve durable local control of their disease and limited risk of toxicity. Further studies will be necessary to determine if there are any reliable predictive factors for radiation-related spinal cord toxicity after SBRT.